To investigate the possibility of using urinary estradiol-17 beta-glucuronide (E2-17G) measured by direct radioimmunoassay to monitor ovulation induction with human menopausal gonadotropin (hMG), serum estrogen and urinary E2-17G levels were determined daily by 21 women treated with hMG for a total of 32 treatment cycles. Urinary E2-17G was measured in 24-hour and overnight specimens. A significant correlation was found between serum estrogens (primarily estradiol) measured by radioimmunoassay without preceding chromatography and urinary E2-17G excretion measured at 24 hours and overnight. The correlation was not significantly improved by correcting the 24-hour and overnight urinary E2-17G excretion levels with creatinine measurements. Although there was significant correlation between serum estrogens and urinary E2-17G measured by direct radioimmunoassay, the urinary E2-17G concentrations observed when serum estrogens levels indicated preovulatory follicle maturation (ie, at serum estrogens levels between 500 and 1000 pg/ml) varied so much that a clinical decision to trigger or not to trigger ovulation with human chorionic gonadotropin could not be reached in each case. These data indicate that significant correlation is not the only prerequisite for a new method to replace a proved procedure. Further studies are required to determine the reliability of monitoring hMG therapy with direct E2-17G radioimmunoassays in overnight urine collections.